Evaluation of a commercially available second-generation immunoglobulin G enzyme immunoassay for detection of Helicobacter pylori infection

Abstract

We evaluated a commercially available second-generation anti-H. pylori immunoglobulin G enzyme immunoassay (EIA) (Cobas Core Anti-Helicobacter pylori EIA; Roche S. A., Basel, Switzerland) for serodiagnosis of H. pylori infection. The results of the assay were assessed in relation to the results of bacterial culture, urease testing, and histological Giemsa stain of gastric biopsy specimens from 1,134 patients with a variety of symptoms relating to the upper gastrointestinal tract. H. pylori was detected in biopsy specimens from 660 (58.2%) patients: 6 had a normal mucosa, 123 had chronic gastritis only, and 531 were found to have chronic active gastritis by histology; endoscopy showed duodenal and gastric ulcers in 137 and 64 patients of the last two groups, respectively. The test was evaluated with different age and ethnic groups. The prevalence, sensitivity, specificity, and positive and negative predictive values were, respectively, (i) for Belgian patients between 18 and 40 years old, 34, 93, 95, 91, and 96%; (ii) for Belgian patients more than 40 years old, 53, 96, 91, 93, and 95%; and (iii) the Mediterranean patients more than 17 years old, 87, 94, 70, 95, and 64%. All sera showing discordant immunoassay results compared with the results of histology and culture of biopsy specimens, as well as those with borderline immunoassay results, were tested further by immunoblotting. Among the EIA results considered false negative, we demonstrated an absence of seroconversion in 14 of 19 patients tested by immunoblotting. Among the EIA results considered false positive, immunoblotting showed the presence of specific antibodies in 28 of 37 patients tested. Among the borderline results obtained in the first assay with 22 patients' sera, a second assay showed positive results in 10 patients (8 were positive by immunoblotting) and negative reactions in 10 patients (9 were negative by immunoblotting), whereas 2 remained borderline. These data indicate that sera showing borderline immunoassay results must be tested again. In conclusion, this commercially available second-generation EIA, which is easy and quick to perform, was found highly reliable for the serodiagnosis of H. pylori infection.